The field of psychiatry has made substantial progress towards understanding mental illness using basic neuroscience methods, but the explanatory gap between cellular hypotheses and clinical phenomena remains vast. This gap is particularly evident in our understanding of schizophrenia, a devastating disorder whose core feature is disrupted cognition. Schizophrenia patients present with debilitating cognitive deficits, not adequately treated by available therapies. Understanding and restoring cognitive function is critical to improving patients' lives. One way to close this gap is to investigate the clinical phenomena by combining several scientific methodologies, at multiple levels of analysis. Therefore, this proposal broadly aims to align functional neuroimaging with biophysically-realistic computational models of neural function and to test model predictions using safe and reversible pharmacological manipulations in healthy volunteers. It further aims to directly compare findings to deficits observed in schizophrenia patients. Ultimately, the current proposal will bridge levels of explanation to mechanistically understand cognitive dysfunction in schizophrenia. One severely compromised cognitive operation in schizophrenia is working memory: the ability to temporarily hold and manipulate information in mind. Disruptions in working memory compromise patients' ability to track thoughts, ideas, and feelings, severely limiting even basic functioning. Although functional neuroimaging studies repeatedly link working memory disturbances to prefrontal dysfunction, synaptic mechanisms remain elusive. One leading hypothesis proposes disruption in the balance of excitation and inhibition in cortical micro-circuitry caused by hypo-function of the N-methyl-D-aspartate (NMDA) glutamate receptor. However, to test this hypothesis in relation to disrupted cognition, and to ultimately develop medications that alleviate cognitive dysfunction in schizophrenia, we need to go a step beyond neuroimaging. We need an understanding of working memory dysfunction at the level of cellular mechanisms, which is where treatments are developed. The specific aims of this proposal are: i) to extend an established biophysically-realistic computational model of working memory to 'mimic' hypothesized NMDA receptor pathology and use it to make behavioral and neural predictions regarding deficits observed in schizophrenia; ii) experimentally test those predictions using a leading safe pharmacological model of schizophrenia that perturbs the precise mechanism in healthy volunteers, namely transient NMDA antagonism via ketamine; iii) to relate these pharmacological results to deficits observed in patients using behavior and functional neuroimaging. The proposed project will close the explanatory gap and help develop a multi-level mechanistic understanding of cognitive dysfunction in schizophrenia. Ultimately, the success of this research will fertilize rationally-guided treatments and improve the lives of people suffering from this devastating disorder.